Mathematical Notation James H. Davenport University of Bath/Chair, - PowerPoint PPT Presentation

Mathematical Notation James H. Davenport University of Bath/Chair, IMU’s Committee on Electronic Information and Communication 3 February 2016 Davenport Mathematical Notation

The lay person’s view difficult “Pour moi, c’est de l’Alg` ebre” is the French for “It’s all Greek to me” universal science fiction stories have humans communicating with aliens via mathematics precise “mathematically precise” is a common phrase unambiguous follows naturally from precise Davenport Mathematical Notation

The mathematician’s view difficult Clearly not, and indeed helpful: [Bou70] the abuses of language without which any mathematical text threatens to become pedantic and even unreadable. universal Pretty much so, though we all have our horror stories precise Well, of course, otherwise we wouldn’t use it. unambiguous follows naturally from precise Davenport Mathematical Notation

If you challenge the mathematician on unambiguity, say with “is (1 , 2) a permutation (in cycle notation) an open interval (at least if you’re anglo-saxon) a (row) vector perhaps you wouldn’t put commas in, but then � 3 � what is ? 2 an ordered pair of integers . . . ”? the response is “it’s unambiguous in my context”: a tactical retreat to local unambiguity Davenport Mathematical Notation

Local unambiguity is very local Consider the following fragments A G i < G ∀ i < n A’ G sub i is a subgroup of G for i less than n B R i < R ∀ i < n B’ R sub i is a subring of R for i less than n C k i < K ∀ i < n C’ k sub i is a subfield of K for i less than n This causes real problems for my colleague who reads examination papers to blind students Davenport Mathematical Notation

“any identifier is as good as any other” is what we all preach, but E = mc 2 is not A = π r 2 and A G i < G ∀ i < n A’ G sub i is a subgroup of G for i less than n B R i < R ∀ i < n B’ R sub i is a subring of R for i less than n C k i < K ∀ i < n C’ k sub i is a subfield of K for i less than n shows that our practice is rather different. [Wat08] shows that each top-level MSC has a unique distribution of the first six identifiers (37 [Dynamical Systems] and 58 [Global Analysis] share n , x , i , k , t ) Davenport Mathematical Notation

Juxtaposition It is normal to say that juxtaposition indicates multiplication (MathML’s symbol &InvisibleTimes; ) or function application (MathML’s &ApplyFunction; ) [Dav08], but in fact the general rules are more complex, and highly context-sensitive. In general, we can state the observed properties of juxtaposition as left right meaning example weight weight normal normal lexical sin normal italic application sin x italic italic multiplication xy (or &InvisibleComma; ) M ij italic normal multiplication a sin x digit digit lexical 42 (or &InvisibleComma; ) M 12 digit italic multiplication 2 x digit normal multiplication 2 sin x Davenport Mathematical Notation

Juxtaposition Table continued normal digit application sin 2 (but note the precedence in 2 sin 3 x ) italic digit error x 2 x 2 or x 2 ? (but reconsider) 4 1 digit fraction addition 2 &InvisiblePlus; application − 1 italic greek a φ (as in group theory) i.e. φ ( a ) italic ( unclear f ( y + z ) or x ( y + z ) what is f ( g + h )? Typography ( \thinspace etc.) can help, but how many document readers (automatic or untrained human) recognise this? Can any-one explain satisfactorily why 2 sin 3 x cos 4 x means 2 · (sin(3 · x )) · (cos(4 · x )), and not, say, 2 · (sin(3 · x · cos(4 · x ))) or 2 · (sin 3) · ( x · cos(4 · x ))? “trigs abhor nesting” isn’t sufficient? Davenport Mathematical Notation

Though this is GDML not ICMI it is worth noting that this overloading of juxtaposition causes real pedagogic problems To illustrate this, I often ask teachers to write 4 x and 4 1 ⁄ 2 . I then ask them what the mathematical operation is between the 4 and the x, which most realize is multiplication. I then ask what the mathematical operation is between the 4 and the 1 ⁄ 2 , which is, of course, addition . . . [Wil11, p. 53]. There was a lengthy debate on LinkedIn in October 2013, around (the Excel evaluations of) 4^3^2 and -3^2 . Note also the MatLab feature that 3i^2 ⇒ − 9 but 3*i^2 ⇒ − 3 The author has seen 1 7 ⁄ 8 mis-OCRed as 17 ⁄ 8 . Davenport Mathematical Notation

Well-enshrined Problems 0 ∈ N ? In theory solved by [ISO], but in practice inconsistent O When we write sin( x ) = O (1) we really mean sin( x ) ∈ O (1) Good A few texts are starting to write ∈ sin 2 When we write sin 2 x we really mean (sin x ) 2 ⑧ and sin − 1 x does not mean (sin x ) − 1 , and sin − 2 x is conflicted Sadly We have to write log log log x because log 3 x is taken M 12 (Entry 1,2 of a matrix, or the 12th Mathieu group?) and the whole &InvisibleComma; mess Davenport Mathematical Notation

Lesser problems: metavariables [AS64, (16.25.1)] defines � u pq 2 ( t ) d t Pq ( u ) = 0 � u sn 2 ( t ) d t , etc.) but p , q ∈ { s , c , n , d } (so Sn ( u ) = 0 except when q is s , when � u � � pq 2 ( t ) − 1 d t − 1 Pq ( u ) = u , t 2 0 Also “So X = ( X , ⊑ X ) for X equal to T , S , V ”, defining T , S , V in one go Davenport Mathematical Notation

Lesser problems: surprise i = 1 : 10 is being used to mean for (i=1;i<=10;i++) , rather than i = { 1 , 2 , . . . , 10 } . D n sometimes means the dihedral group with n elements, sometimes the group on n points (with 2 n elements) ± Is used in many different ways: tan z 1 ± tan z 2 = sin( z 1 ± z 2 ) , [AS64, (4.3.38)] cos z 1 cos z 2 is shorthand for two equations, but [AS64, Equations 4.6.26,27] � � � � 1 + z 2 1 + z 2 Arcsinh z 1 ± Arcsinh z 2 = Arcsinh z 1 2 ± z 2 1 � � � ( z 2 1 − 1)( z 2 Arccosh z 1 ± Arccosh z 2 = Arccosh z 1 z 2 ± 2 − 1) is “every value of the left-hand side is a value of the right-hand side and vice-versa” Davenport Mathematical Notation

Recommendations in increasing order of boldness(!) ISO Standards exist: the community should follow them And de facto standards such as [CCN + 85, Nat10] Paper is no longer a scarce resource: some space-saving techniques (metavariables, ± ) are actually counterproductive Typesetting is easy, so sin 2 x is not cheaper than (sin x ) 2 , and is notationally polluting ∈ should be used if that’s what we mean Juxtaposition should be used more sparingly, and properly annotated in MathML Davenport Mathematical Notation

Bibliography I M. Abramowitz and I. Stegun. Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables, 9th printing. US Government Printing Office , 1964. N. Bourbaki. Th´ eorie des Ensembles. Diffusion C.C.L.S. , 1970. J.H. Conway, R.T. Curtis, S.P. Norton, R.A. Parker, and R.A. Wilson. Atlas of finite simple groups . Clarendon, Oxford, 1985. Davenport Mathematical Notation

Bibliography II J.H. Davenport. Artificial Intelligence Meets Natural Typography. In S. Autexier et al. , editor, Proceedings AISC/Calculemus/MKM 2008 , pages 53–60, 2008. ISO. International standard ISO 31-11: Quantities and units — Part 11: Mathematical signs and symbols for use in the physical sciences and technology . International Organization for Standardization, Geneva. National Institute for Standards and Technology. The NIST Digital Library of Mathematical Functions. http://dlmf.nist.gov , 2010. Davenport Mathematical Notation

Bibliography III S.M. Watt. Mathematical Document Classification via Symbol Frequency Analysis. In S. Autexier et al. , editor, Proceedings AISC/Calculemus/MKM 2008 , pages 29–40, 2008. D. William. Embedded Formative Assessment. Solution Tree , 2011. Davenport Mathematical Notation